ChatGPT sparked public interest in the topic of artificial intelligence (AI) a year and a half ago, and most people now wonder, “When will AI bring a significant change to my life?”
This question resonated especially during the big tech’s annual developer conference season, which started in mid-May. This is the time when tech companies set up booths and strive to captivate customers with their visions for the immediate future of technology.
![](/wp-resource/resources/articles/2024/06/ai-biggest-promise-and-two-problems-to-solve-01.png)
Visions and Reality of Tech Companies
While the advent of ChatGPT may have sparked public imagination, most people’s interest was limited to entering questions into this text-based chatbot. Since then, the tech industry’s focus has largely shifted from the uses of generative AI to the race to build the infrastructure needed to deliver it at a mass scale.
Headlines in the media have been dominated by news of more powerful large language models, massive spending on powerful new chips, and the proliferation of giant, power-hungry data centers needed for AI processing. Now, these powerful technical capabilities are getting closer to actual tech users. Microsoft’s biggest news this month is the announcement of next-generation AI-powered PCs branded as Copilot+ coming this year, powerful enough to handle AI without calling a remote data center.
In the process, Microsoft has challenged Apple, claiming its new PCs will surpass Apple’s MacBook. The AI race is now heating up in the world of personal computing and smartphones.
However, none of this answers the most important question for most consumers: “How (and when) will this expensive new technology improve my life?”
Current Applications and Limitations of AI
So far, generative AI has led to the spread of text boxes online that provide answers to questions (including services like Meta’s WhatsApp and Instagram), services that help with writing emails or documents, and various services that summarize text blocks, such as web digests that Google has begun offering at the top of search results. It is still unclear how many people actually use these features.
As this month’s events have shown, tech companies have far bigger ambitions. They aim for personal digital assistants that can predict user needs and mediate much of their online activity, and even further, digital agents that can take actions on behalf of users.
These ideas were at the core of Google’s event two weeks ago, Microsoft’s event last week, and the announcement of OpenAI’s new model GPT-4o. However, if this is AI’s greatest promise, it remains just that—a promise.
Two Problems to Solve
For AI to truly transform our lives, two fundamental issues must be resolved.
One is enabling AI models trained on past data to understand and appropriately respond to new situations. As Demis Hassabis, head of Google’s AI research division, says, AI must “be able to understand and respond to the world as complex and dynamic as humans do.” This is a challenging task. It’s not just about avoiding “hallucinations” or occasional glaring mistakes AI systems are prone to. It also means consistently providing genuinely useful results by fully understanding the context.
Google claims significant progress in this area with its latest Gemini model, which includes an expanded “context window” to help the system maintain awareness of complex situations. However, much remains to be proven for technology to match human understanding of the world.
Another related issue is making AI communication with humans as natural as talking to a person. According to those building the systems, only then will this technology fulfill its potential.
Microsoft’s CEO Satya Nadella stated that to achieve this, “we need to learn how to make computers understand us, not us learning to understand computers.” Despite his claim that this goal is within reach, Demis Hassabis and others warn that attempts to achieve “natural” interaction with computers are still a “very high bar.”
OpenAI offered a glimpse of what might be possible in the future with a demonstration of its AI model GPT-4o, designed to operate in an informal conversational style. Yet, there remains a significant gap between staged demos and effective real-world products. Predicting when AI will make a significant breakthrough in the consumer world is still difficult.
Conclusion: Expectations for AI’s Future
We are filled with anticipation about how AI will change our daily lives. However, many problems remain to be solved. For AI to truly improve our lives, we must wait for technological advancements that help it understand the world like humans and communicate naturally.
As we look forward to the future of AI that will change our lives, we must watch for the positive changes that technology will bring. For these changes, we must welcome technological progress and believe in its possibilities.
Reference: Financial Times, “AI’s biggest promise for consumers remains just that — a promise”