Is Google taking AI too far?

Sundar Wednes

writer icon Paige Whitehead     Google   |   Tech Ethics     🕐 20. Jun. 2018

Google Duplex sounds human, and can make telephone calls for you. While some have applauded the technology, others voiced serious concerns over the ethics of a robot that mimics human voices. 

Automated Assistants
With Google Duplex, users can let Google schedule their appointments, with the aim of saving time that would otherwise be spent on the telephone. The way Google sees it, everyone could have the benefit of a personal assistant right on their phones and computers.

Assisting small businesses

Small businesses who do not have an online presence could benefit from the services Google Duplex offers. Google intends to update business information that is discovered during the exchange while the artificial intelligence (AI) is making calls

For example, the AI could call to find out about opening times, and then update the information online on behalf of the company. This could benefit both the businesses and the public by having information more readily available. 

It is not clear whether Google plans to call businesses for updates themselves, or simply wait for one of their users to call and then update the information as it is acquired. If it is the latter, this implies that the content of the calls will be seamlessly uploaded to the internet at the AI’s discretion. Many would see this as an invasion of their privacy.

Helping people with disabilities
Google Duplex could also benefit people with disabilities. A person who has a speech impediment, phone anxiety, or is in some way unable to use the phone in a conventional way, could use the technology to take over those responsibilities, while they go about their daily lives. 

Google Duplex aims to offer a variety of assistance for a variety of needs. It is not the assistance offered that is causing concern. It is the way they plan to go about it.

Ethical Concerns
Google Duplex was unveiled During Google's developer conference in May 2018, as Google's newest AI technology. A snippet of a real conversation between the AI and unsuspecting businesses were shown. The humans on the receiving end of the call appeared to have no idea that they were talking to a robot. This raised a multitude of concerns regarding the ethics around speaking unknowingly to a robot.

Google has since stated that the snippets were recorded in early development and that Google Duplex will now introduce itself as an automated assistant on the phone. 

Imitating real people

Google also displayed their abilities to recreate existing human voices for their automated assistant. They used the voice of singer John Legend to demonstrate this new technology.

Not only can the AI mimic a human voice, but it can completely recreate the voice of an existing person. This could lead to a number of issues, and some people were shocked that Google had launched this technology without simultaneously providing any sort of guideline or plan for how to avoid others abusing it. 

Risk of abuse

One might wonder if anyone could potentially get a sample of someone else's voice and use it to create an AI capable of imitating just about anybody. This could lead to someone sabotaging another person by using their voice to make incriminating phone calls or to mislead others. 

Many people want to know what precautions Google have put in place to prevent others from misusing Duplex. For example, if someone were to make an imitation of a politician's voice for their AI assistant, they could potentially use the voice to impersonate the politician and make complications on a national, even international level.

Google's AI Principles

One month later, on 7th June 2018, Google published their AI principles. This explained how they will use the technology, as well as what they will not do. Many see this as a direct response to the criticism levied at them following the unveiling of Google Duplex.

Google and the Military
Google has already been under scrutiny following their involvement in the U.S. government program Project Maven. Project Maven involved the use of Google's AI technology, and could have potentially facilitated drone strikes. Google employees fought against their own company and succeeded, causing Google to decide not to renew their contract. 

Though Google has promised not to renew their contract regarding Project Maven, their AI principles state that they will continue to work with the military. With an aim to "augment the critical work of these organizations", the only promise Google has made is that they will no longer work on developing weapons.

A question of privacy

With a company as influential as Google, more scrutiny should be given to their privacy terms and conditions. Especially if they are working on AI technology and accepting contracts from the military. With the recent controversy concerning Facebook’s invasion of privacy, Google and the public need to be aware of the possibilities. 

The Facebook scandal was a huge invasion of privacy. Yet, people are usually more aware of what they do and do not disclose in a social media setting. Are people monitoring themselves as closely when it comes to a Google search? 

One wonders what information Google keeps on its users, as well as the conceivability of the US government, or an outside source, gaining access to that information. If artificial intelligence is updating business hours on Google based on user activities, who and what else could be being updated?

Holding Google Accountable

Recently, Google has presented a profound ignorance when it comes to the public’s expectations. They did not appear to anticipate the backlash they experienced following the demonstration of Google Duplex.

Additionally, they did not take their employees into consideration when signing military contracts. And the wording on their AI principles appears vague enough to get them out of a lawsuit. 

With so many ethical concerns, it is incredible that Google did not anticipate the outcry over ethics when creating an AI that mimics a human voice. Their principles on AI use were not released until a month later, following the objections. 

Did Google get too swept up in their own technology to not realise the ramifications? These are questions that Google should clarify in more than a single blog post.

Google holds the lives, and privacy, of over one billion people in their hands. One hopes they remember that their primary concerns are people, not AI.

We believe that information should be free and will therefore never put up a paywall.

If you like reading our reports about the Scandinavian business scene and would like to donate towards the upkeep of the site, we would be very grateful. Click here to donate.

Most Popular Articles of November

Most Popular Articles of this Year