5. Discussion
This study contributes a new finding to our understanding of
entrepreneurial pitching strategy. That is, this study demonstrates that
the current level of AI-generated avatar technology can be a
competitively and equitably viable video spokesperson consideration for
entrepreneurs. This is a valuable revelation, given the importance of
video for entrepreneurial outreach like on crowdfunding platforms
(Carradini & Fleischmann, 2023; Courtney et al., 2017; Johan & Zhang,
2020; X. Li et al., 2016; Mollick, 2014). This study also thus adds to
our understanding of digital oratory’s evolution, particularly as it
relates to the subgenre of entrepreneurial digital oratory.
At a fundamental level, this study confirms the implicitly and
explicitly held understanding of digital oratory – that delivery and
narrator persona, even in the digital context, matter even beyond the
technical or logical argumentative content. As research has indicated
prior to this study and then confirmed through this study’s variables,
the entrepreneurial spokesperson can absolutely have an impact on the
success or failure of a pitch. This study advances our general
understanding that in digital oratory, the orator matters, and places
and confirming that understanding squarely within a commercial context.
The prudent entrepreneur, then, would ensure that the nonverbal delivery
in their digital oratory-based proposal is performed in such a way that
it serves as a valuable vehicle for gaining engagement, buy-in, and
goodwill. What that data here suggests is that AI-generated avatars are
now worthy of consideration for the entrepreneur seeking to meet those
goals.
Because of the viability of avatars demonstrated here, our understanding
of digital oratory as a genre of speaking, for private, public, and
commercial purposes, must now include a further complication –
awareness that this form of “embodied” communication can now
effectively be organic or synthetic. While extensive exploration have
confirmed the limitations synthetic characters have due to the uncanny
valley phenomenon (Gray & Wegner, 2012; MacDorman & Entezari, 2015;
Tinwell et al., 2013; Tondu, 2015), this new era of avatars appear to
have moved beyond that limiting feature for the majority of respondents.
The increased effectiveness of synthetic speakers may have been achieved
by the type of avatars used in this system, fashioned off from live
human video models instead of pure computer-generated animation. That
situates the avatar much closer already to the expected narrators in the
broader, accepted digital oratory genre. The AI-powered dynamism coded
into the avatar’s seemingly authentic delivery style then allows it to
be believable for the majority of viewers, especially those unaware of
the backend reality. That synthetic delivery prowess is what appears to
allow this class of synthetic spokespersons to be equally as
persuasively successful. This understanding is consistent with the
existing literature on the impact of persona and delivery in
entrepreneurial pitch success (Allison et al., 2022; Davis et al., 2017;
Jiang et al., 2019; Y. Li et al., 2021).
Of course, the results here to not offer a blanket endorsement of
avatars as human spokesperson replacements. Quite the opposite. The very
design of the experiment, controlling the setting and gestures of the
live actors to match the digital puppets’ limited options, points to the
freedom and flexibility that organic human actors provide. However, the
time and cost saving, along with the professional aesthetic so easily
achieved, will have to factor against those creative freedoms. With the
avatar programs also allowing users to upload video of themselves to
create their own bespoke custom avatars, even the positive
self-promotion tactics previously found valuable (Korzynski et al.,
2021) are an option now in the digital context, further complicating the
entrepreneur’s decision calculus.
Entrepreneurs considering an avatar speaking for their endeavor should
also carefully weigh the remaining risks of the uncanny valley, even if
just caused by poor execution in this evolving technology. When the
synthetic nature of the digital spokesperson was identified, the
respondent ratings were less positive. In the case where the viewership
was expected to be skeptical, when the creator was not familiar with the
technology, or when there was a small sample size of viewers, the risks
of a diminished score may rise beyond the liking of an entrepreneur.
This is not to say that the risks of avatar-awareness negate the value
of the AI avatars. The data here suggests otherwise quite directly as
well. Not only is the fake highly effective, with most respondents not
identifying the synthetic avatar as such, the overall average was
equally as high, regardless of which video type was viewed. Thus, for an
entrepreneur or other persuasively minded creator, as more and more
viewers see their pitch, the likelihood increases that they too will not
have anomalous negative reactions, but instead will also produce the
same equivalent results. In the context of crowdfunding, for example,
when the viewership can be in the tens of thousands to even millions,
the data here suggest that there is no statistical risk in using an AI
avatar. Even further, for those wishing to reduce risks caused by ethnic
or gender inequalities in entrepreneurship(Chang & Xu, 2023), the
synthetic spokesperson of a varying demographic may be an appealing,
even if ethically complicated, consideration.
Even further, because the delivery of the avatar is already highly
effective at fooling more than half of viewers, and because significant
financial investments and user data is being collected in massive
quantities, the coding needed to further refine the animation is almost
guaranteed to be provided to the companies creating this software, like
Synthesia and its competitors. This all but guarantees that the risks
directly associated with identifiability in the avatar will continue to
decrease.
It should also be noted that this study was also based on a single
context and single script. The skilled pitch writer has many other
alternative ways of creating value or even additional risk even with a
constrained use of the avatars. As the landscape of digital oratory
undergoes its current seismic shifts, the conventions of the genre will
thus change as new rhetorical and videographic approaches are tested and
normed. For example, it may or may not be advantageous to transparently
admit that the spokesperson is an AI avatar. Future studies, as well as
industry use-case exemplars, will help scholars and practitioners better
understand how these new contextual best practices develop.