House of Lords launches inquiry into Large Language Models

The name “Aviva Investors” as used in this material refers to the global organization of affiliated asset management businesses operating under the Aviva Investors name. Each Aviva investors’ affiliate is a subsidiary of Aviva plc, a publicly- traded multi-national financial services company genrative ai headquartered in the United Kingdom. Except where stated as otherwise, the source of all information is Aviva Investors Global Services Limited (AIGSL). They should not be viewed as indicating any guarantee of return from an investment managed by Aviva Investors nor as advice of any nature.

• Ethical risks – Generative content could produce harmful, biased, or misleading messaging without oversight and governance.
• From what we’ve seen, Llama 2 is more than powerful enough for users to start exploring the technology in a safe way.
• Whether released via API or open source, a single issue with a model at the foundation stage could create a cascading effect that causes problems for all subsequent downstream users.
• Data leakage occurs when sensitive information, proprietary algorithms, or confidential details are unintentionally exposed through the responses of a Language Model (LM).

It uses natural language processing, machine learning and data analytics to automate and enhance various aspects of legal work. We have seen its extraordinary power, and you’d be hard-pressed to find a lawyer who’s seen it in action who hasn’t also been impressed. For the PoC, I feel it’s better to build a generic and easily adaptable chatbot that uses public or easily accessible services to provide the functionality. As this is a generic solution, it can easily be extended to use different services to create a custom-tailored genrative ai chatbot or data analysis tool for any specific task. “Over the last decade, the Darktrace Cyber AI Research Center has championed the responsible development and deployment of a variety of different AI techniques, including our unique Self-Learning AI and proprietary large language models. We’re excited to continue putting the latest innovations in the hands of our customers globally so that they can protect themselves against the cyber disruptions that continue to create chaos around the world,” added Stockdale.

Generative AI: eight questions that developers and users need to ask

From another perspective, it also leaves the opportunity to harness the potential of generative AI to chance. This book will provide you with insights into the inner workings of the LLMs and guide you through creating your own language models. You’ll start with an introduction to the field of generative AI, helping you understand how these models are trained to generate new data. Next, you’ll explore use cases where ChatGPT can boost productivity and enhance creativity. You’ll learn how to get the best from your ChatGPT interactions by improving your prompt design and leveraging zero, one, and few-shots learning capabilities. The use cases are divided into clusters of marketers, researchers, and developers, which will help you apply what you learn in this book to your own challenges faster.

Such occurrences can result in unauthorised access to sensitive data, privacy infringements, and security breaches. Inadequate filtering of sensitive information, overfitting or memorisation of confidential data during the LM’s training, and misinterpretation or errors in the LM’s responses are some of the factors that can contribute to data leakage. This article is part of our ‘New Beings’ series, examining the role of AI in development. We’re finding the output from Llama 2 to be extremely high quality, with meaningful and correct content. We’ve also found Llama 2 to have good reasoning ability, meaning it’s very difficult to fool it with trick questions. Obviously running the 70bn-parameter model takes more compute power, but the result is a very sophisticated, clever model.

EU data protection authorities have also recently started to look at some generative AI providers. In March 2023, the Italian data protection authority (Garante) blocked ChatGPT’s processing of personal data (effectively blocking the service in Italy) until ChatGPT complies with certain remediations required by the authority. In April 2023, the Spanish data protection authority (AEPD) initiated its own investigation. It is likely other data protection authorities will follow – the European Data Protection Board (EDPB) has since launched a task force on ChatGPT. European data protection authorities are concerned with the use of personal data in AI systems, including to train it, and questions around lawful processing, transparency, data subject rights and data minimisation in particular.

Other companies may want to buy some breathing space to better assess and understand the risks and formulate better informed policy or guidelines. Foundation models can be made available to downstream users and developers through different types of hosting and sharing. The AI products we use operate within a complex supply chain, which refers to the people, processes and institutions that are involved in their creation and deployment. For example, AI systems are trained using data that has been collected ‘upstream’ in a supply chain (sometimes by the same developer of the AI system, other times by a third party. Artificial General Intelligence (AGI) and ‘strong’ AI are sometimes used interchangeably to refer to AI systems that are capable of any task a human could undertake, and more. This is partly because they are futuristic terms that describe an aspirational rather than a current AI capability – they don’t yet exist – and partly because they are inconsistently defined by major technology companies and researchers who use this term.

Legal

Since ChatGPT’s release in November 2022, generative AI has entered public discourse across the world. According to GlobalData, over a million social media posts about artificial intelligence (AI) have been made across Twitter and Reddit in the last year. At the international level, G7 leaders recently announced the development of tools for trustworthy AI through multi-stakeholder international organisations through the ‚Hiroshima AI process‘ by the end of the year. In addition, Senate Majority Leader Chuck Schumer has announced an early-stage legislative proposal aimed at advancing and regulating American AI technology.

Firstly, you’re sending corporate data to a centralised cloud, potentially leaving you open to loss of IP. The second (and in some cases, even greater) risk is what the service can learn about you from your interactions with it. For those looking to use generative AI for business, we think these risks far outweigh the benefits of trying something new. Rather than generating search results in a list based on keywords, DeepSights provides business users with complete, natural language answers to their market and consumer intelligence questions.

Yakov Livshits

Get Technical Training

Foundation models require an extremely large corpus of training data, and acquiring that data is a significant undertaking. That data is cleaned and processed, sometimes by the company that develops the model, other times by another company. Once an AI model is put into service, it may be relied on by ‘downstream’ developers, deployers and users, who use the model or build their own applications on it. NVIDIA NeMo enables organizations to build custom large language models (LLMs) from scratch, customize pretrained models, and deploy them at scale. Included with NVIDIA AI Enterprise, NeMo includes training and inferencing frameworks, guardrailing toolkits, data curation tools, and pretrained models. Unlike generic models like GPT, Observe.AI’s proprietary large language model (LLM) is trained on a domain-specific dataset of hundreds of millions of customer interactions.

From now on, consider the main user interface of every digital tool to be your native tongue. Just like in science fiction, you’re free to lay back in your chair, and have the bot do your bidding. Especially in product- or software-related translations, the variability aspect might lead to a major quality drop and quickly extend to other parts of the business.

This enables it to support a diverse set of AI-based tasks that are highly specific to contact center teams. Our ability to gather market information across broad datasets, bring it into uniform decision-making platforms and compare it against benchmarks will improve thanks to AI. It could help improve risk management and enhance the customer experience by providing access to a wealth of real-time data.

Ask them a question, and LLMs will use a huge corpus of human generated text (the internet), shove it into a neural network model (a so-called transformer), and spit out the most reasonable-sounding answer. “At Darktrace, we have long believed that AI is one of the most exciting technological opportunities of our time. With today’s announcement, we are providing our customers with the ability to quickly understand and control the use of these AI tools within their organizations. But it is not just the good guys watching these innovations with interest – AI is also a powerful tool to create even more nuanced and effective cyber-attacks. Society should be able to take advantage of these incredible new tools for good, but also be equipped to stay one step ahead of attackers in the emerging age of defensive AI tools versus offensive AI attacks,” concluded Gustafsson. These new risk and compliance models for Darktrace DETECT and RESPOND make it easier for customers to put guardrails in place to monitor, and when necessary, respond to activity and connections to generative AI and large language model (LLM) tools.

Delivering the AI + Data + CRM program

Auto-GPT can also apparently be used to improve itself – its creator says it can create, evaluate, review and test updates to its own code that can potentially make it more capable and efficient. Auto-GPT is also able to help businesses to autonomously increase their net worth by examining their processes and making intelligent recommendations and insights about how they could be improved. Immediately, the generative image AI creates a super-realistic movable image of Angelina Jolie, Scarlett Johanssen, Zendaya or whoever you like. Then, using an LLM engine you can start to ask questions to your ‘DreamLover’, and she will respond thanks to the LLM.

These training steps can be repeated and improved upon to refine the accuracy and effectiveness of the GPT over time. They can write essays, create poetry, generate conversational agents, translate languages, and even mimic specific writing styles. These models are being used in various applications like chatbots, content creation, and educational tools, making human-like text generation more accessible and efficient. We are here to support organisations, enabling them to scale and maintain public trust. Our recently updated Guidance on AI and Data Protection provides a roadmap to data protection compliance for developers and users of generative AI. Our accompanying risk toolkit helps organisations looking to identify and mitigate data protection risks.

In just a short period, we will likely see massive changes in how customers find products, engage with companies and experience brands. Generative AI can create personalized, easy-to-understand communications at scale, to keep citizens informed about public sector initiatives and services. By making complex information more accessible, LLMs can foster increased public engagement and trust in government institutions. Large Language Models and generative AI outcomes can be insightful, interesting, and extremely simple to understand by business users with varying degrees of comfort with technology and visualizations.