Is Google trying too hard?

If you're a developer — or not — you must be familiar with Google's annual conference, Google I/O. It's exciting, to say the least. We get to hear about the latest innovations from one of the world's most innovative tech giants. The new releases, be it from Google or Microsoft are sure to get plenty in livestreams of late. And this is certainly not because they maintain an element of suspense and are on the verge of introducing something truly revolutionary for the tech world. Google I/O, without any surprise, was all about AI.

But this year's event left many people scratching their heads. Google was rather like a chirpy child keen to show his or her latest idea of brilliance — it kept mentioning “Help me write” as if to portray it as a lifesaver of sorts. It is designed to help users write more professional-sounding emails and messages. On the face of it, this sounds great! How often do we sit at our desks, wondering how to respond to emails that ask us to perform out of the bounds of our job definitions? We might have to resort to loewhaley’s reels to find corporate subtleties for our feelings; to contain them into language that might fit the appropriateness of the office culture. So, yes, a tool that can design our emails with care can never be shooed away. Or, can it?

The tool falls short of its intended goal. In fact, the AI-generated messages created by "Help me write" often sound robotic — nuance is a no-show here. And no one’s saying this just out of spite. It can be derived from the demo Google had on display. The tool generated a job application response that sounded like it was written by a generic corporate drone, rather than a real person. There’s very little that such generated emails can reveal about the personas that act behind the email. It isn’t a bitter “everything must be ‘human’” narrative that’s working on this logic. It’s just that we can use the best of both worlds.

When people jibe at writers or those who are involved in design and content creation, citing how AI has started nagging at the ends of their job profiles, there’s a key fact that they overlook: anything that has been trained to reproduce results will need a honing of these results. So, when you run content through AI and expect it to rephrase it or rewrite it for you — or provide details about someone and ask it to create a message directed at them — you need to add an extra eye for detail. You must make sure that this is indeed the tone and message you wish to send across. And once you’ve added that human touch from your side, go on, and click send. It’s quite easy to confuse messaging for a blandly transactional act.

That’s not all.

Google also announced the Search Generative Experience which is supposed to radically change the way search results appear. And perhaps help us find information better? SGE uses advanced AI models to summarize and combine information from all over the web and presents it in a human-like format, instead of the usual “ten blue links” we’re used to seeing. So, you would have a box that pops up first with all relevant information, thereby, in most cases, ending your search journey then and there. As expected, not everyone is happy about this quick fix. Publishers are worried that SGE will hurt their traffic by keeping users on Google.com instead of sending them to their sites. There’s also concern that Google’s AI may be lifting content without giving proper attribution, which has been a long-standing issue between Google and publishers.

Barry Diller, Chairman of IAC, which owns websites like All Recipes and The Daily Beast, is leading a group of publishers that are exploring ways to prevent AI firms like Google from scraping their content. They believe that current “fair use” restrictions need to be redefined, and that copyright law may need to be changed. But, Google continues to defend its role as the protector of the online search ecosystem — SGE is nothing more than an experiment, they say, perhaps as an anticipatory bail. They still refrain from telling the public about the training sources used.

This is just another chapter in the whole AI tale. But it’s safe to say that Google is stretching itself and playing a game that comes off too hard on itself. The sudden surge in demand for Bing — a couple of months back — still seats itself as a sparkling, distasteful memory for Alphabet. Pichai and Co. should probably dwell on taking things a little slow, don’t you think? That’s at least better than having allegations thrown around as in the case of Bard.