Opening the Gates on AI
Any new technology that’s so disruptive is bound to make people uneasy, and that’s certainly true with artificial intelligence.
Bill's Recent Thoughts on AI
Bill Gates has seen two demonstrations of technology that struck him as revolutionary.
- 1980 - when he was introduced to the graphical user interface
- 2022 - when he witnessed OpenAI's ability to train an artificial intelligence that could ace an advanced placement biology exam.
The recent demonstration sparked him to think about what AI will achieve over the next five to 10 years.
His mind naturally gravitated to areas that his philanthropy, the Gates Foundation, is already focused on and how the advancements in AI can help them in their mission. Below is a summary of his points sprinkled with my thoughts.
Productivity Enhancement: Push vs. Pull
The professional landscape is going to be more drastically affected than most people are aware of or can even comprehend at this point. Building on the potential of AI, Gates envisions a future where personal agents can drastically enhance productivity by handling various tasks. This idea is based on recent advancements in AI, such as those seen with GPT. In my opinion, he more accurately depicts it as having a white-collar worker available to help you with various tasks. I think this drastically undersells it.
Microsoft describes this personal agent currently as having a co-pilot in its initial product roll-outs. This could be due in part to it still requiring a lot of hand-holding from humans to get helpful outputs but also it sounds less frightening when the AI is working with you. Initial products that have been popping up like weeds during the summertime still require a basic understanding of prompt engineering. Effective prompt writing is when the given instructions are informative enough to in turn get a response that meets your needs. The more specific you can be generally the better output you will receive. As humans advance these technologies around core models we should see more of a pull flow.
Push example: Give me a five-bullet outline on how to improve our sales funnel based on the following input [input info here].
Pull example: Please find a five-bullet outline on how to improve your sales funnel based on it being listed as a task you already have on your calendar today and also leveraging yesterday's meeting notes on the topic.
Gates later comments that the advancements will allow workers to improve their tasks and free up their time for work they'd prefer to spend their time on. It should allow people to free themselves from mundane, repetitive tasks, and focus more on core function improvements to the business.
Looking beyond individual personal agents, there is potential for company-wide agents to revolutionize how businesses operate. It will be interesting to see how these agents are leveraged in the near future. Microsoft is already launching products to meet this perceived need with Microsoft 365 having AI chat built in. Having access to all or a large portion of your company's internal data will be an evolution of search+ that will completely upend how to educate ourselves on actions going on more broadly in the company.
As with any new technology, some risks and problems must be addressed. In the case of AI, one concern is how workers will need to be retrained to adapt to the changing landscape. However, there are even more pressing issues to consider, such as the risk of incorrect answers being provided and the potential for strong AIs to conflict with humanity's interests. Governments will need to help people with this transition. AI itself might be able to aid in the process.
He has a couple of comments throughout the paper to help reassure people that AI will not completely replace them but in the specific areas he mentions I am not convinced.
The rise of AI will free people up to do things that software never will—teaching, caring for patients, and supporting the elderly, for example.
In AI's current state, yes, this is true. However, when you combine advancements in AI with robotics though this goes out the window. Plenty of people already are self-taught in the world finding little to no value in some of the degrees they attain for crushing amounts of debt. Patients are cared for by people but robots can be programmed to do that eventually and also won't be irritated by your repetitive questions after a 12-hour shift. Caring for the elderly is extremely costly both from a monetary perspective and a time perspective. A co-pilot in this situation could allow for a more enjoyable, engaging living experience that could be difficult for humans to offer without high labor costs.
Health
Gates' vision for AI in healthcare is ambitious, but not without its challenges. While AI has the potential to increase access to care in underserved areas and improve the efficiency of healthcare delivery, there are also concerns about data privacy, the accuracy of diagnoses, and the potential for biases in the training data. Gates acknowledges the need for rigorous testing and regulation to ensure that health AIs are beneficial overall. However, he also sees the potential for AI to rapidly improve the quality of care for those who currently have limited access to it.
Gates states that people will need to see evidence that health AIs are beneficial overall and that they'll need to go through robust testing with proper regulation being implemented. He also mentions this will lead to longer adoption curves.
However, I believe that people who currently have limited access to care will likely embrace AI quickly, even without extensive testing and regulation. For example, in areas where there is only one doctor for a village of 500 people, AI could drastically improve access to care and reduce the workload of the doctor. While it is important to ensure that health AIs are effective and safe, we should not delay their implementation in areas where they could save lives and improve health outcomes. Additionally, while Gates emphasizes the need for AI to be rigorously tested, I believe that AI could be adopted more quickly if people are able to see evidence of its benefits firsthand. A careful balance will need to take place to make sure care reaches people as soon as they need it but in as safe of a manner as possible.
Education
AI has the potential to revolutionize not only healthcare but also education. As technology continues to evolve and enable people to self-educate, we may see a profound impact on education that we haven't seen in decades according to Gates. In this section, we'll explore some of Bill Gates' thoughts on the role of AI in education.
Gates' Education Takeaways
- Computers have not had a meaningful effect on students' achievement in education so far.
- AI-driven software will revolutionize the way people teach and learn in the next 5 to 10 years.
- AIs can assist teachers and administrators in assessing a student's understanding and giving advice on career planning.
- AIs will need diverse training data sets to be unbiased and reflect the different cultures where they will be used.
- Educators are already discussing ways to adapt to the new technology, and conversations will continue for quite some time, with some teachers finding clever ways to incorporate AI into their work.
Individual learning applications are some of the most powerful use cases that AI will allow for in and outside of the traditional classroom. As AI models continue to advance and increase their personalization to their students it is not far-fetched that you will graduate from grade to grade with your agent. Learning outcomes can be individually tailored to the type of learner.
Four main types of learners:
- Visual learners
- Auditory learners
- Kinesthetic learners
- Reading/writing learners
AI will need to continue to evolve to see this level of personalization come to fruition but we seem to be moving in that direction. Areas that I believe are less clear than Gates points out are around the great relationships still being needed between teachers and students.
Even once the technology is perfected, learning will still depend on great relationships between students and teachers. It will enhance—but never replace—the work that students and teachers do together in the classroom.
It's extremely hard to make a claim and back it up using the word "never" when dealing with new technology. Relationships between great teachers and students will still have profound effects even with the advancements in AI but they first need to be present for that to happen. If artificial general intelligence is fully evolved to master any human task then it will by definition be able to replace the work that students and teachers do together in the classroom. A less harsh perspective would be that it would allow for at least another alternative similar to having multiple teachers at your disposal vs. being stuck with only one for an entire year before graduating to a new teacher and a new teaching style.
In a lot of situations across the United States teachers are overworked, poorly paid, and have no incentive to go above and beyond what is needed in their day-to-day. A lot of teachers are driven by a personal passion to make a difference but depending on classroom sizes, available resources, and family issues with students that get brought into the classroom, it can be very difficult to always offer a great learning experience. It becomes even harder to have a personalized catered learning environment. The structure of existing classrooms is to move the collective pool of students forward very rarely adapting to learning styles or fast-paced learners.
Risks and problems with AI
More immediate risks stem from incorrect answers being provided and the known issue of hallucinations in popular AI chatbots like ChatGPT. Gates hits on the eventual evolution of "strong" AIs that will be able to more effectively self-teach and set their own goals which could conflict with humanity's interests. Despite these concerns, Gates believes that we are only scratching the surface of what AI can accomplish. While strong AI is still a long way off, there is no doubt that AI will continue to shape our world in profound ways.
The Next Frontiers
Gates acknowledges to the reader that AI will continue to be a topic of conversation for the foreseeable future and offers his thoughts on what can help guide these conversations.
Gates' suggested principles to help guide conversations around AI:
- The conversation around AI should balance fears about its downsides with its ability to improve people's lives.
- Market forces won't naturally produce AI products and services that help the poorest, and governments and philanthropy can ensure that AIs are used to reduce inequity, not increase it.
- We're only at the beginning of what AI can accomplish and existing limitations today will be overcome quickly.
The ability to have balanced insightful conversations will be hindered by a general lack of understanding of this technology. This gap of knowledge can be further amplified by having material implications if policies are drafted by politicians that don't understand what its capabilities truly are both positive and negative. With the speed that technology is moving, I'm not sure policies can realistically keep pace or be forward-thinking enough to be equally effective five to 10 years from now. Our political system much like our education system was built for a different era. Eventually, it will need an upgrade to keep up with modern times.
Will AI help?
Conclusion
Bill Gates' insights into the potential of AI have given us a glimpse into a future that is both exciting and uncertain. As with any new technology, there are risks and challenges that must be addressed, but the benefits could be transformative for society. It's up to us to have informed conversations and shape the development of AI in a way that ensures it benefits everyone, not just the privileged few.
If you're interested in reading Gates' original article it's linked below. Subscribe below to be kept up to date on my thoughts about AI and the interesting tools being built around it.
Bill Gates