Why Agentic AI Will Soon Make ChatGPT Look Like A Simple Calculator

Lawsuit Alleges Microsoft Trained AI on Private LinkedIn Messages

LinkedIn acknowledges that it uses personal data and creative content for AI training and will share that data with third parties for model training. But the lawsuit raises the question of whether LinkedIn has included the contents of private InMail messages available to paying customers in the personal data disclosed. The fair use doctrine was designed for specific, limited scenarios—not for the large-scale, automated consumption of copyrighted material by generative AI.

• They must use the muscles in their legs to push up against this mysterious unseen force.
• Importantly, attorneys are expected to understand potential risks such as hallucinations, biased outputs, and the limitations of GAI’s ability to understand context.
• The Opinion also addresses the emerging question of when GAI use should be disclosed to clients or courts.
• The key distinction between generative and agentic AI lies in their approach to tasks and decision-making.

Generative AI has emerged as a pivotal tool in enhancing cyber security strategies, enabling more efficient and proactive threat detection and response mechanisms. As the shortage of advanced security personnel becomes a global issue, the use of generative AI in security operations is becoming essential. For instance, generative AI aids in the automatic generation of investigation queries during threat hunting and reduces false positives in security incident detection, thereby assisting security operations center (SOC) analysts[2]. Generative AI models are trained on vast datasets, often containing copyrighted materials scraped from the internet, including books, articles, music and art. These models don’t explicitly store this content but learn patterns and structures, enabling them to generate outputs that may closely mimic or resemble the training data. Questions about decision-making transparency, ethical boundaries, and appropriate levels of autonomy need careful consideration.

Future Prospects

Maybe the AI must have a “body” or at least a semblance of embodiment to fully grasp how to operate in the physical world. Physical AI will be the make-or-break of whether those mechanizations are compatible with humans and operate properly in the real world or instead are endangering and harmful. Aha, the AI is doing its calculations and responses based on data and not on any real-world first-hand experience. They must use the muscles in their legs to push up against this mysterious unseen force.

To ensure generative AI serves society without undermining creators, we need new legal and ethical frameworks that address these challenges head-on. Only by evolving beyond traditional fair use can we strike a balance between innovation and protecting the rights of those who fuel creativity. Most datasets used to train generative AI models include copyrighted materials without the creators’ consent. Creators have the right to control how their work is used, and the absence of their consent undermines ethical and legal defenses.

Whether analyzing network traffic for anomalies or identifying phishing attempts through advanced natural language processing (NLP), LLMs have proven to be invaluable tools. Security firms worldwide have successfully implemented generative AI to create effective cybersecurity strategies. An example is SentinelOne’s AI platform, Purple AI, which synthesizes threat intelligence and contextual insights to simplify complex investigation procedures[9]. Such applications underscore the transformative potential of generative AI in modern cyber defense strategies, providing both new challenges and opportunities for security professionals to address the evolving threat landscape. Generative AI technologies are transforming the field of cybersecurity by providing sophisticated tools for threat detection and analysis. These technologies often rely on models such as generative adversarial networks (GANs) and artificial neural networks (ANNs), which have shown considerable success in identifying and responding to cyber threats.

Attorneys should understand both the capabilities and limitations of specific GAI technologies they employ, either through direct knowledge or by consulting with qualified experts. This is not a one-time obligation; given the rapid evolution of GAI tools, technological competence requires ongoing vigilance about benefits and risks. The Opinion suggests several practical ways to maintain this competence, including reading about legal-specific GAI tools, attending relevant continuing legal education programs, and consulting with technology experts. Importantly, attorneys are expected to understand potential risks such as hallucinations, biased outputs, and the limitations of GAI’s ability to understand context. By simulating phishing scenarios and generating tailored educational materials, these models help organizations improve their employees’ ability to recognize and respond to cyber threats.

Beyond these technical domains, the report reveals an intriguing mix of human capabilities rising in importance, with risk mitigation, assertiveness, and stakeholder communication all featuring prominently. Moreover, a thematic analysis based on the NIST cybersecurity framework has been conducted to classify AI use cases, demonstrating the diverse applications of AI in cybersecurity contexts[15]. As it continuously learns from data, it evolves to meet new threats, ensuring that detection mechanisms stay ahead of potential attackers [3]. This proactive approach significantly reduces the risk of breaches and minimizes the impact of those that do occur, providing detailed insights into threat vectors and attack strategies [3].

How scalable solutions are helping federal agencies unlock AI’s potential

As these systems become more sophisticated and widespread, they have the potential to transform industries, enhance human capabilities, and open new frontiers in human-machine collaboration. The key will be ensuring that we develop and deploy these technologies thoughtfully, with clear frameworks for accountability and control. The study calls for a multi-faceted approach to enhance the integration of LLMs into cybersecurity. Developing comprehensive, high-quality datasets tailored to cybersecurity applications is essential to improve model training and evaluation.

The computational demands of large models often strain resources, making scalability a critical concern, especially in real-time operational environments. Additionally, the lack of high-quality, domain-specific datasets hampers the ability to fine-tune models effectively. Security professionals need to trust and understand model-generated recommendations to act on them confidently, necessitating improvements in explainability and transparency. Another case study focuses on the integration of generative AI into cybersecurity frameworks to improve the identification and prevention of cyber intrusions. This approach often involves the use of neural networks and supervised learning techniques, which are essential for training algorithms to recognize patterns indicative of cyber threats. However, the application of neural networks also introduces challenges, such as the need for explainability and control over algorithmic decisions[14][1].

Generative AI has emerged as a transformative force in technology, creating text, art, music and code that can rival human efforts. However, its rise has sparked significant debates around copyright law, particularly regarding the concept of fair use. If you are going to learn AI, there are a number of free classes online that would be a great place to start. AI is here to stay, but it won’t be replacing humans anytime soon, as the human touch still needs to be added to any AI content. People who know how to use AI will replace those who are not trained or certified in AI. The notion of agentic AI is that we could have multiple generative AI instances serving as your agents or assistants to accomplish some particular tasks.

• AI-generated text might reorganize or paraphrase existing content without offering unique insights or value. While these factors have worked well in traditional scenarios like criticism, parody or education, generative AI presents unique challenges that stretch these boundaries. This statistic underscores a fundamental shift in how organizations view talent and potential.

Survey: College students enjoy using generative AI tutor – Inside Higher Ed

Survey: College students enjoy using generative AI tutor.

Posted: Wed, 22 Jan 2025 08:01:50 GMT [source]

Addressing these challenges requires proactive measures, including AI ethics reviews and robust data governance policies[12]. Collaboration between technologists, legal experts, and policymakers is essential to develop effective legal and ethical frameworks that can keep pace with the rapid advancements in AI technology[12]. The Opinion also addresses the emerging question of when GAI use should be disclosed to clients or courts. While not every use of GAI requires disclosure, attorneys must inform clients when GAI outputs will influence significant decisions in the representation or when use of GAI tools could affect the basis for billing. For court submissions, attorneys must carefully verify GAI-generated content, including legal citations and analysis, to meet their duties of candor toward tribunals under Rule 3.3.

Sure, a human can show another human how to jump in the air and do the splits, but it won’t especially sink in until the person being shown the demonstration attempts the physical act themselves. Microsoft is one of the biggest investors and developers in the AI space, but it’s not the only one—see the others on our list of the top AI companies to better understand who is defining this dynamic technology. One of the most significant fair use factors is the effect on the market for the original work. Generative AI threatens to disrupt creative markets by producing high-quality content at scale. AI lacks the intent to create something transformative, making it challenging to meet this critical fair use requirement.

By recognizing subtle indicators of malicious activities, such as unusual network traffic or phishing attempts, these models can significantly reduce the time it takes to detect and respond to cyberattacks. This capability not only prevents potential damages but also allows organizations to proactively strengthen their security posture. In a broader context, generative AI can enhance resource management within organizations.

During cybersecurity incidents, LLMs provide detailed analyses, suggest mitigation strategies, and, in some cases, automate responses entirely. This level of automation enables cybersecurity professionals to concentrate on addressing complex threats. Prompt injection attacks are particularly concerning, as they exploit models by crafting deceptive inputs that manipulate responses. Adversarial instructions also present risks, guiding LLMs to generate outputs that could inadvertently assist attackers.

How generative AI is paving the way for transformative federal operations

With new features and tools being released on a consistent basis, it can be difficult for professionals to know where to start or how to keep up in a constantly changing field. Below, 20 Forbes Business Council members share tips to help professionals effectively break into the AI or generative AI field of work. The pilots targeted diverse use cases, including officer training, semantic search for investigative data and hazard mitigation. “These pilots taught us valuable lessons about responsible AI use, governance and measuring success,” says Kraft. DHS has summarized those insights in its newly released DHS GenAI Public Sector Playbook.

Generative AI offers significant advantages in the realm of cybersecurity, primarily due to its capability to rapidly process and analyze vast amounts of data, thereby speeding up incident response times. Elie Bursztein from Google and DeepMind highlighted that generative AI could potentially model incidents or produce near real-time incident reports, drastically improving response rates to cyber threats[4]. This efficiency allows organizations to detect threats with the same speed and sophistication as the attackers, ultimately enhancing their security posture[4]. Generative AI technologies utilizing natural language processing (NLP) allow analysts to ask complex questions regarding threats and adversary behavior, returning rapid and accurate responses[4]. These AI models, such as those hosted on platforms like Google Cloud AI, provide natural language summaries and insights, offering recommended actions against detected threats[4].

Another notable aspect is the Opinion’s treatment of different types of GAI tools and required validation. Tools specifically designed for legal practice may require less independent verification compared to general-purpose AI tools, though attorneys remain fully responsible for all work product. The appropriate level of verification depends on factors such as the tool’s track record, the specific task, and its significance to the overall representation. Boilerplate consent provisions in engagement letters are deemed insufficient; instead, lawyers should provide specific information about the risks and benefits of using particular GAI tools. As agentic AI systems become more sophisticated, we’re likely to see a fundamental shift in how we interact with artificial intelligence. Rather than simply issuing commands and receiving outputs, we’ll develop more collaborative relationships with AI systems that can engage in genuine back-and-forth dialogue, propose alternative solutions, and even challenge our assumptions when appropriate.

By automating routine security tasks, it frees cybersecurity teams to tackle more complex challenges, optimizing resource allocation [3]. Generative AI also provides advanced training environments by offering realistic and dynamic scenarios, which enhance the decision-making skills of IT security professionals [3]. Looking ahead, the prospects for generative AI in cybersecurity are promising, with ongoing advancements expected to further enhance threat detection capabilities and automate security operations. Companies and security firms worldwide are investing in this technology to streamline security protocols, improve response times, and bolster their defenses against emerging threats. As the field continues to evolve, it will be crucial to balance the transformative potential of generative AI with appropriate oversight and regulation to mitigate risks and maximize its benefits [7][8].

Research into lightweight architectures and parameter-efficient fine-tuning techniques can address scalability issues, enabling broader adoption. Generative AI, while offering promising capabilities for enhancing cybersecurity, also presents several challenges and limitations. One major issue is the potential for these systems to produce inaccurate or misleading information, a phenomenon known as hallucinations[2]. This not only undermines the reliability of AI-generated content but also poses significant risks when such content is used for critical security applications.

Challenges and Limitations

Perhaps the AI will by text-basis logic assume that if the dog is dropped, it will bounce like a rubber ball. I’m assuming that if I asked you the same question, your answer would be about the same. You see, since the days of being a baby and a toddler, you eventually figured out that dropping a rubber ball from shoulder height will fall to the ground and then bounce. Well, let’s go ahead and ask ChatGPT a question involving the physical action of dropping a rubber ball and find out what the AI has to say.

For legal practitioners engaged in technology law and policy, the Report serves as a comprehensive reference for understanding both current regulatory frameworks and potential future developments in AI governance. Each section includes specific recommendations that could inform future legislation or regulation, while the extensive appendices provide valuable context for interpreting these recommendations within existing legal frameworks. The Report emphasizes the need for balanced, sector-specific approaches to AI regulation that promote innovation while protecting against potential harms, with particular attention to ensuring equitable access and protecting consumer rights across all sectors. Beyond examining these key guidelines, we’ll also explore practical strategies for staying informed about AI developments in the legal field without becoming overwhelmed by the rapid pace of change. Whether you’re just beginning to explore AI tools or are already integrating them into your practice, understanding these guidelines is crucial for maintaining professional standards and maximizing the benefits of these transformative technologies. The shift from purely generative to more agentic AI represents a fundamental reimagining of what artificial intelligence can be.

As Maggioncalda points out, this global disparity in skills adoption could reshape how organizations think about talent acquisition and development. As emerging markets demonstrate increasing proficiency in AI skills, companies are likely to tap into these new talent pools, potentially altering traditional hiring patterns and creating more globally distributed teams. In an era where AI capabilities are expanding exponentially, the ability to communicate effectively, show assertiveness, and manage stakeholder relationships has become more crucial than ever. The rise in demand for these skills suggests that while AI may handle many tactical tasks, strategic thinking and relationship building remain uniquely human domains. Among the evaluated models, GPT-4 and GPT-4-turbo achieved top accuracy scores, excelling in both small-scale and large-scale testing scenarios. Meanwhile, smaller models like Falcon2-11B proved to be resource-efficient alternatives for targeted tasks, maintaining competitive accuracy without the extensive computational demands of larger models.

It could be that thinking is not separable from our senses such as having ears, eyes, noses, taste, and touch. Without those bodily capabilities, it could be that a brain and mind would not ultimately formulate into a thinking capacity. A brain and mind might be an empty vessel without having had the experiences of sensory inputs from the likes of their body working in physical environments. Sci-fi plot lines have often delved into this devilish riddle by having brains floating in vats and disconnected from an actual body. Reality and physical movement are like the air we breathe; it is all around us and we conventionally take it for granted.

Over half of executives believe that generative AI aids in better allocation of resources, capacity, talent, or skills, which is essential for maintaining robust cybersecurity operations[4]. Despite its powerful capabilities, it’s crucial to employ generative AI to augment, rather than replace, human oversight, ensuring that its deployment aligns with ethical standards and company values [5]. The application of generative AI in cybersecurity is further complicated by issues of bias and discrimination, as the models are trained on datasets that may perpetuate existing prejudices. This raises concerns about the fairness and impartiality of AI-generated outputs, particularly in security contexts where accuracy is critical.

Generative Adversarial Networks (GANs)

These advanced technologies demonstrate the powerful potential of generative AI to not only enhance existing cybersecurity measures but also to adapt to and anticipate the evolving landscape of cyber threats. ANNs are widely used machine learning methods that have been particularly effective in detecting malware and other cybersecurity threats. The backpropagation algorithm is the most frequent learning technique employed for supervised learning with ANNs, allowing the model to improve its accuracy over time by adjusting weights based on error rates[6]. However, implementing ANNs in intrusion detection does present certain challenges, though performance can be enhanced with continued research and development [7].

Fine-tuned models consistently outperformed general-purpose ones, demonstrating the importance of domain-specific customization. The Bipartisan House Task Force on Artificial Intelligence (“Task Force”), established in February 2024, represents a significant legislative initiative to comprehensively examine AI’s impact across American industries and institutions. The Report’s structure reflects a methodical analysis of AI’s implications across multiple sectors, with each section providing sector-specific findings and actionable recommendations.

• The financial services section details how AI is reshaping traditional banking and financial operations, with recommendations for maintaining consumer protections while fostering innovation.
• While not every use of GAI requires disclosure, attorneys must inform clients when GAI outputs will influence significant decisions in the representation or when use of GAI tools could affect the basis for billing.
• Generative AI is revolutionizing the field of cybersecurity by providing advanced tools for threat detection, analysis, and response, thus significantly enhancing the ability of organizations to safeguard their digital assets.
• This is not a one-time obligation; given the rapid evolution of GAI tools, technological competence requires ongoing vigilance about benefits and risks.

While this is hugely impressive, these systems are essentially reactive; they respond to specific prompts without any real understanding of context or long-term objectives. To counter these challenges, the study emphasizes the importance of robust input validation techniques. Advanced adversarial training can help models identify and resist malicious inputs, while secure deployment architectures ensure that the infrastructure supporting LLMs is resilient against external threats. These strategies collectively enhance the integrity and reliability of LLM applications in cybersecurity.

GANs are also being leveraged for asymmetric cryptographic functions within the Internet of Things (IoT), enhancing the security and privacy of these networks[8]. Under Rules 5.1 and 5.3, managerial attorneys must establish clear policies governing the firm’s permissible use of GAI, while supervisory attorneys must ensure both lawyers and non-lawyer staff comply with professional obligations when using these tools. This includes implementing comprehensive training programs covering GAI technology basics, tool capabilities and limitations, ethical considerations, and best practices for data security and confidentiality. The Opinion also extends supervisory obligations to outside vendors providing GAI services, requiring due diligence on their security protocols, hiring practices, and conflict checking systems. The ability of LLMs to analyze patterns and detect anomalies in vast datasets makes them highly effective for identifying cyber threats.

Their ability to correlate diverse data points allows for more comprehensive investigations, which not only aid in recovering from incidents but also provide insights to prevent future breaches. This capability makes LLMs an essential tool in the forensic analysis of sophisticated cyberattacks. There are also concerns regarding bias and discrimination embedded in generative AI systems. The data used to train these models can perpetuate existing biases, raising questions about the trustworthiness and interpretability of the outputs [5].

The movement toward more agentic capabilities may be accelerating, with recent reports suggesting various AI labs are exploring ambitious new directions. According to Bloomberg reports, OpenAI has been rumored to be working on a project codenamed «Operator,» which could potentially enable autonomous AI agents to control computers independently. We’re beginning to see the first signs of convergence between generative and agentic capabilities in mainstream AI tools. OpenAI’s recent introduction of scheduled tasks in ChatGPT represents an early step in this direction.

Generative AI models are trained on massive datasets, often containing millions of works. While individual pieces may contribute minimally, the sheer scale of usage complicates the argument for fair use. Fair use traditionally applies to specific, limited uses—not wholesale ingestion of copyrighted content on a global scale. The study highlights LLMs’ applications across domains such as malware detection, intrusion response, software engineering, and even security protocol verification. Techniques like Retrieval-Augmented Generation (RAG), Quantized Low-Rank Adaptation (QLoRA), and Half-Quadratic Quantization (HQQ) are explored as methods to enhance real-time responses to cybersecurity incidents.

How do we ensure these systems remain aligned with human values and interests while maintaining their ability to operate independently? How do we balance the benefits of increased automation with the need for human oversight and control? These are critical questions that will shape the future development of agentic AI systems. EWeek has the latest technology news and analysis, buying guides, and product reviews for IT professionals and technology buyers.

Another major vulnerability is data poisoning, where malicious actors inject false or misleading data during the training phase, compromising the reliability of the model. Denial-of-service (DDoS) threats further exacerbate these issues by overwhelming LLM-based systems with excessive requests, rendering them inoperable during critical moments. GANs play a crucial role in simulating cyberattacks and defensive strategies, thus providing a dynamic approach to cybersecurity [3]. By producing new data instances that resemble real-world datasets, GANs enable cybersecurity systems to rapidly adapt to emerging threats. This adaptability is crucial for identifying subtle patterns of malicious activity that might evade traditional detection methods [3].

The realistic scenarios created by LLMs enhance the effectiveness of training initiatives, fostering a culture of security awareness within organizations. Generative AI has revolutionized incident response by automating routine cybersecurity tasks. Processes such as patch management, vulnerability assessments, and compliance checks can now be handled with minimal human intervention.