Google unveils next steps in AI technology
At its annual developer conference, Google showed off new products that make use of artificial intelligence. A handful of firms are using AI to reshape our world. Is this a good thing?
Ever spotted a flower and wondered what it is? Soon enough, you will be able to ask your phone: Google is developing an app that scans images and recognises the objects within them.
“Google Lens” is just one of the new uses of artificial intelligence (AI) that the company unveiled at its conference this week. Other products included an improved personal assistant and a tool that automatically groups your photos into themed albums. This was “some of the best practical applications of AI… I’ve seen”, gushed the BBC’s tech correspondent.
The conference was a reminder of just how integral AI is becoming to our lives. It helps shape political campaigns, run our health care and organise our Facebook news feeds. The technology is also being put to more ambitious uses. Tesla is building self-driving cars; Google is trying to teach computers to understand human languages and even create art.
In the past, such large-scale scientific projects would have been the government’s business. Today, advances in AI are largely being driven by a small number of mega-rich companies. In 2017, five of these companies are set to spend a whopping $60 billion on research and development, much of which will involve AI.
Last October, the Obama administration recommended a huge increase in public spending on AI. But Donald Trump has shown little interest in boosting funding for science. Many are worried. AI can be a force for good — and for bad. The technology is too risky, they say, to be left in the hands of private firms.
To be fair, the firms share some of these concerns. Last year they teamed up to launch Partnership on AI, an organisation that educates the public about AI and encourages transparency in the industry. Google and others “open-source” some of their research (ie, make it public) and regularly publish papers.
Yet all of this is voluntary. Unlike governments, companies are not obliged to work in the public interest, and they are driven by profit. They are good at making gadgets — but should they be trusted with the future of AI?
Too clever by half
Relax, say some. The companies have resurrected the field of AI after decades of underfunding, and they understand its benefits and risks better than anyone. Bad AI is bad for all humanity, so it is in their interest to use it well. Their transparency initiatives prove that they are serious about this.
That’s so naive, reply others. AI is a hugely complex field with vast implications for society. The government has a duty to participate in anything this important. It must ensure that the technology is being used fairly and carefully. Companies can say all they want, but we can never hold them to account for their actions.
- How do you feel about the rise of AI?
- Are governments necessarily more trustworthy than private firms?
- Watch the BBC’s video presenting Google’s latest products in Become An Expert. As a class, rank them in terms of how exciting they are.
- Prepare and give a three-minute presentation arguing that AI is either a force for good or for bad. After everybody has gone, count: are there more optimists or pessimists in the class?
Some People Say...
“Real stupidity beats artificial intelligence every time.”— Terry Pratchett
What do you think?
Q & A
- What do we know?
- What is certain is that AI will become an ever greater part of our lives. It could improve many things, like the efficiency of our health care and the safety of our roads (thanks to self-driving cars). But it also has the potential to do damage, such as by taking over jobs in services or manufacturing, or through its use in sophisticated new weaponry.
- What do we not know?
- To state the obvious: we have little idea how AI will evolve in the future. Optimists say that AI, like previous technological revolutions, will create as many jobs as it destroys, and that nations will work together to regulate its military uses. Pessimists argue that humans simply cannot evolve as fast as AI. The worst-case scenario, depicted in many sci-fi films, is that computers begin to control us.
- Google Lens
- The app’s release date is not yet confirmed.
- Personal assistant
- Billed as a “smart speaker”, Google Home performs various functions — music playback, weather reports — which you activate with voice commands.
- Create art
- You can check out an AI-generated song in Become An Expert.
- Amazon, Apple, Facebook, Microsoft and Alphabet (Google’s parent company) — otherwise known as the “Frightful Five”.
- $60 billion
- By comparison, in 2015 the US government spent $67 billion on scientific research unrelated to defence. A mere $1.1 billion of this went toward AI projects.
- Much of which
- Last year, Google’s chief executive declared that the company would henceforth be “AI-first”.
- Donald Trump
- The president’s budget blueprint, released in March, proposes cuts to a wide range of scientific bodies. It is worth noting that this blueprint only sets out Trump’s priorities — Congress passes the final budget.
- In 2015, a group of big players in Silicon Valley launched OpenAI, a non-profit AI firm committed to making its research public. See Wired’s article in Become An Expert.