Is it OK to fake empathy?

Late last year I was part of a university presentation and group discussion about the impressive progress of AIs (in this case GPT) in lots of fields. But something was raised in that conversation that I cannot get out of my mind.

The conversation turned to GPT’s capacity to carry out very human-like conversations and its exciting potential applications in care, coaching and counselling spaces. This included its ability to show, or fake, empathy.

I raised my hand and asked …

“I am doing work trying to find ways to reduce social isolation in aged care by giving older people and their carers novel ways to promote meaningful interactions. Empathy is part of that. Is it OK to fake empathy?”

The speaker answered, “Yes, if it helps, I think it is.”

Six months later, and with a whole lot more improvement in AI’s ability to do this, I find myself constantly returning to the central dilemma of that conversation: is it really OK to allow an AI (from a text-based chatbot to a photo-realistic avatar or household robot) to fake empathy in a conversation with a human?

What do you think?

I don’t think it is.

For a start, I think most people would say they do not feel good about the idea of being on the receiving end of fake empathy from another human. It just kind of, feels wrong. But it’s more than a feeling. Faking empathy breaks down trust, it can be seen as manipulative, even narcissistic, and likely lots of other not good things.

We all know this. Authenticity is a key, perhaps the key, foundation of meaningful relationship and genuine connection. It certainly appears that when people perceive inauthenticity the positive effects that empathy has on trust are suppressed. We don’t like to be tricked in relationships.

So, how can anyone be actively pursuing the goal of faking empathy?

Well, perhaps because, like lots of things about relationships, showing real empathy takes time, effort, understanding, vulnerability, and for professional encounters – a skilled approach.

In care settings, such commitment is not always possible, or at least gets given the back seat, due to such things as: insufficient staff availability, or skill level, or discomfort, or unwillingness from a person to participate at a prescribed time. So an AI able to show AE (artificial empathy, yes, that’s really a term) has obvious potential. Especially when it could be offering it to thousands of people at the same time and relieve carers and family members from the communication burden. (I don’t of course really feel it is a burden, but it can be perceived that way).

Same goes for coaching, counselling, career guidance, interviewing, not to mention – just outright companionship, friendship, love, marriage?!

A key underpinning here is highlighted by one more thing the speaker at the university meeting added to his answer to my question.

The speaker added,

“… and when it gets so good that the person being cared for cannot tell the difference, why does it matter?”

Which of course just doubles down on the whole thing. Now we have the compound dilemma: Do we tell them? If we don’t, they continue to feel empathy. If we do, we risk eroding the trust and even cause damage. Sounds a lot like that swallow the spider to catch the fly thing.

But I think there’s another point here too. I don’t think it’s just about the recipient of the empathy. Surely empathetic connection is two-way and both people feel and gain something. So, while we may “relieve” ourselves and our businesses of some of the financial and commitment burdens required by the need to be truly present for those we care for or supervise, we may well end up losing a massive dose of vital human connection. And everybody loses.

So, I for one, will be doing my best to foster and participate in genuine, honest, authentic, empathetic, connection. At home, with family, with friends, with clients, with new people I meet. It may be less efficient and exciting to shareholders than the burgeoning artificial alternative, but perhaps that’s the point – it’s human to human. And, or course, sometimes it might include a little assistance from a little pack of CCS Cards.

I like Sherry Turkle’s request for developers in the last paragraph of her MIT Tech Review article, The chatbot I’ve loved to hate, written in 2020 following isolation and all of us needing to use Zoom:

“Some were tempted to talk to bots. But when we had all the time in the world to be with our machines, most of all, we missed each other. We wanted to reach past technology to the full embrace of the human. We suffered when our families and friends got sick alone, had babies alone, had too many dinners alone, and indeed died alone.

Engineers: make a better Zoom. Make better tools for us to be together when alone. There is no need to compete with the empathy that defines what is unique about being a person.”