The Next Frontier of AI: From Chatbots to Courtrooms and Beyond

Artificial intelligence (AI) is changing the way we communicate, work, and learn. Chatbots like ChatGPT can generate realistic and engaging texts on various topics and even converse with humans in natural language. These chatbots are powered by large language models (LLM) that can process and learn from huge amounts of data.

But chatbots are just the tip of the iceberg. AI is also advancing in other domains, such as law, science, engineering, and robotics. Fabian Stephany and Johann Laux recently outlined four examples of developing AI possibilities:

• AI legal help: A startup called DoNotPay claims to have built a legal chatbot that can advise defendants in court. The chatbot can listen to the proceedings and whisper legal arguments to the defendant via an earpiece. However, the chatbot faces legal challenges and ethical questions, as it may not be authorized to practice law or represent humans in court.
• AI scientific support: Scientists are using AI to analyze data and propose novel hypotheses. AI can help scientists discover new patterns, insights, and solutions that may not be obvious to human intuition or bias. For example, a machine learning system ranked chemical combinations for battery materials, saving time and resources for researchers.
• AutoGPT: The next generation of chatbots will be based on GPT-4, a more advanced LLM that can handle different types of data, such as images and speech. These chatbots will be able to generate multimodal content, such as captions, summaries, and translations. They will also be able to understand the context and intent of the user better. But there is more: AutoGPT, an advanced AI tool released by Significant Gravitas, can perform complex tasks by itself, without human input. AutoGPT can split a general goal into sub-tasks and complete them using AI agents that make decisions based on rules and goals. For example, AutoGPT can plan a birthday party by booking a venue, sending invitations, and ordering a cake.
• Humanoid robots: Robots that look and act like humans are becoming more realistic and capable. They can perform tasks that require dexterity, coordination, and social skills, such as serving customers, delivering packages, or assisting elderly people. Humanoid robots have improved significantly since the first Darpa Robotics Challenge in 2015 when many robots struggled to perform basic tasks. Now, startups are developing humanoid robots for warehouse work, using AI fields such as computer vision and power-dense batteries.

These are some of the AI technologies that could shape our future. They have the potential to improve our lives, but they also raise important questions about their feasibility, viability, and acceptability. We should be prepared to embrace them, but also to regulate them.