General Purpose AI: What is it and what can it do?
General purpose AI, often referred to as artificial general intelligence (AGI), is a type of AI capable of performing any intellectual task that a human can do. Unlike narrow AI, which is designed for specific applications such as image recognition or language translation, general purpose AI aims to understand, learn, and apply knowledge across a wide range of domains. This encompasses reasoning, problem-solving, and the ability to adapt to new situations without human intervention. The pursuit of AGI raises significant questions about its potential impact on society, ethics, and the nature of intelligence itself. In this blog we will focus on how it’s current capabilities and how it can contribute to social work.
What can it do?
Understanding and measuring the capabilities of general-purpose AI is crucial for assessing their risks. Existing governance frameworks and commitments rely on precisely measuring general-purpose AI capabilities, but they are a moving target and difficult to measure and define.
Most experts agree that general-purpose AI systems are capable of tasks including:
Assisting programmers and performing small- to medium-sized software engineering tasks.
Creating images that can be hard to distinguish from real photographs.
Engaging in fluent conversation in many languages.
Finding and summarising information relevant to a question or problem from many data sources.
Working simultaneously with multiple ‘modalities’ such as text, video, and speech.
Solving textbook mathematics and science problems at up to a graduate level.
What can’t it do?
Current general-purpose AI systems have an uneven set of capabilities, and still have many limitations. Most experts agree that general-purpose AI is currently not capable of tasks including:
Performing useful robotic tasks such as household work.
Consistently avoiding false statements.
Independently executing long projects, such as multi-day programming or research projects.
AI Agents
AI Agents are a type of general-purpose AI that can make plans to achieve goals, adaptively perform tasks involving multiple steps and uncertain outcomes along the way, and interact with its environment – for example by creating files, taking actions on the web, or delegating tasks to other agents – with little to no human oversight.
General-purpose AI agents can increasingly act and plan autonomously by controlling computers. Leading AI companies are making large investments in AI agents because they are expected to be economically valuable. There is rapid progress on tests related to web browsing, coding, and research tasks, though current AI agents still struggle with work that requires many steps. Many users report AI Agents to be glitchy in real world scenarios.
Getting smarter but still prone to errors
During 2024, general-purpose AI systems have markedly improved at tests of scientific reasoning and programming. These improvements come in part from techniques that let general-purpose AI break down complex problems into smaller steps, by writing so-called ‘chains of thought’, before solving them.
While general-purpose language models are getting smarter and can correctly answer many common-sense and factual questions, they can be inconsistent and make trivial errors.
General-purpose AI systems can achieve performance similar to or better than human experts on some self-contained knowledge and reasoning tasks, but they still make mistakes on easy problems in ways that humans do not.
Opportunities for social work
Reducing Administrative Burden and Managing Workload
General purpose AI can automate routine tasks like data entry, scheduling, and report generation, significantly reducing administrative burdens. Tools like Magic Notes and Diona can record conversations and generate assessments, while AI-powered virtual assistants manage emails and reminders, allowing social workers to focus more on people accessing services.
Making Better Use of Information
AI can analyse both structured and unstructured data to provide holistic insights. It can generate case summaries and analyses from a child's record, enabling social workers to make more informed decisions and provide more personalised support.
Enhanced Search and Retrieval
AI's advanced search capabilities allow social workers to quickly retrieve information from case notes, forms, and documents using intuitive natural language queries. This reduces the time spent searching for information and improves efficiency.
Improving Accessibility
AI-driven apps and portals enable children and families to contribute to their own records using speech-to-text or chatbot prompts, bridging language and learning barriers. Virtual assistants and chatbots help navigate complex systems and provide accessible information, enhancing support for children and families.
Transforming Case Recording
AI can transform case recording by creating more inclusive and representative records. AI-driven tools can generate life story work, mapping, and quality assurance outputs, ensuring accurate documentation of significant events and relationships.
Risks and ethical factors
While AI offers significant opportunities to enhance social care these benefits must be balanced with with careful consideration of risks and ethical factors.
AI systems can inadvertently perpetuate biases present in historical data, leading to unfair outcomes. Privacy and data protection are paramount, especially when dealing with sensitive information about children and families. Ensuring transparency in AI use and obtaining informed consent from those affected are essential to maintaining trust. It is important that people that use services are aware of how their information will be used and have the opportunity to raise concerns if they think their information has been misused or misrepresented by AI or AI developers. By implementing robust ethical guidelines and continuous oversight, the social care sector can harness the power of AI responsibly and equitably.
Conclusion
General purpose AI, or artificial general intelligence (AGI), holds the potential to revolutionise social work reducing administrative burden and managing workload. However, the integration of AI in social care must be approached with caution, balancing the opportunities with the associated risks and ethical considerations. Addressing risks and ethical issues is important for building trust and safeguarding the interests of those served by social care systems. By adhering to robust ethical guidelines and maintaining continuous oversight, the social care sector can responsibly harness the power of AI, ultimately enhancing the quality and effectiveness of services provided to children and families.