Steve Hamilton, Digital Media Director
The Internet is a fascinating place, ripe with possibility, opportunity and anonymity. Businesses of every size and scope have harnessed the power of the web, social media and increased connectivity to grow brand loyalty exponentially. On the other hand, these tools, if implemented poorly, can leave you with egg that is hard to wash off.
Use of innovative and new technology should be tempered. Even the best of tech’s titans can struggle to avoid implementation mistakes. The best lesson of all, don’t leave your reputation in the hands of Internet users, writ-large.
Microsoft, in an attempt to connect with a younger demographic and demonstrate development of its artificial intelligence product, released “Tay.” She was released into the Twitter verse. This fun-loving, innocent AI chatbot was not ready for the reality of the Internet.
That shouldn’t come as a surprise to anyone who has looked at the comments section of online newspapers, but apparently the engineers behind this publicity effort overlooked what could happen.
A day after she was incepted, users had already turned the AI into a parade of the ribald, vulgar and offensive. By giving the users the ability to shape her personality and content, Microsoft essentially set itself up for embarrassment. They relinquished any and all control they had in their own product, and had to watch in horror as users took delight in steering poor Tay into all sorts of trouble.
Avoiding the self-inflicted wound is a fundamental tool in any crisis management situation. In this case, the team behind Tay overlooked an obvious shortcoming. Anyone with an axe to grind, or interested in making Microsoft look bad, could and would encourage its proprietary AI to say awful things.
Microsoft was quick to delete the offending tweets and issue a public statement/apology, but it was too late. Users had already saved and archived a sort of “best-of”, even creating a top Reddit group called “Tay Tweets.” (Some of these are highly offensive and have been included for posterities sake.)
The lesson to be learned here is that users don’t care about what’s best for business, or what’s best for your brand. They could be in it for a laugh. Maybe they don’t like your product, or their family works at a competitor. If you can’t steer your message, or at the very least guide it, it can go into some dangerous territory when your reputation is left open to the Internet.
The second takeaway from this experiment is that once something has been published to the general public online, it needs to be considered as permanent. Even if you delete it, it’s going to exist forever. People take screenshots and the like, you can never fully “erase” a mistake.
Microsoft’s overall brand won’t be tarnished by this episode. It’s likely they won’t lose many contracts or shareholder value because of this specific event. But every business can learn from this event, which involved artificial intelligence.
You should always be in charge of your company’s reputation. If you have a social media account and don’t manage it, you can be made to look foolish. If you create an AI, you better make sure it can filter out hateful speech.
Allow people to participate, but keep the keys as to who gets published. Mashable has a great example with their Vine challenges. Every week they give out a challenge for people to do on Vine by creating a themed 7 second video, and then Mashable chooses the best to feature on their page. Their essentially outsourcing content, but they get to choose what they get behind rather than just throwing it all out there. They choose what’s best for branding, and make sure they grab nothing offensive.
For the time being at least, communications professionals need to keep tabs on the machines, before they take over everything.